GeomGCL

GeomGCL is an implementation of a method for predicting molecular properties using geometric graph contrastive learning. It includes preprocessing steps for molecular datasets and provides a framework for training models on these datasets.